Post on 12-May-2018
1
BIOMECHANICAL EFFECTS OF LAMINOPLASTY AND LAMINECTOMY ON THE
STABILITY OF CERVICAL SPINE
INTRODUCTION
Relevant Anatomy: The spinal column is made up of series of motion segments, each of
which contains an intervertebral disc and its two adjacent facet joints. The resultant spinal canal
houses the spinal cord and nerve roots, while the structural arrangement is well suited to weight-
bearing and diverse head movements (e.g., rotate side to side, flexion and extension). The
cervical spine is made up of the first seven vertebrae in the spine. A typical vertebra consists of
an anterior body and a posterior ring made of articular, transverse, and spinous processes. The
cervical spine also features a complex arrangement of ligaments to supplement its structure and
mobility. Figure 1 shows the general anatomy of a cervical vertebra.
Cervical spinal stenosis is the most common spinal cord disorder in persons more than 55
years of age in North America and perhaps in the world [1]. It is a chronic degenerative
condition of the cervical spine that results in the reduction of spinal canal diameter and thereby
compresses the spinal cord and the associated nerve roots. Surgical options can be broadly
classified into two categories namely, anterior and posterior approaches. This study focuses on
posterior based approach (i.e. laminectomy or laminoplasty) which is considered when multiple
levels of the spine have to be decompressed or when most of the cord compression results from
posterior pathological conditions. Historically, laminectomy has been regarded as the standard
treatment for multi-level cervical stenosis and myeloradiculopathy. However, the results of the
procedure were universally unsuccessful with complications such as segmental instability and
postoperative kyphosis [2]. Because of such concerns, laminoplasty was developed as an
alternative to laminectomy. It is mainly intended to relieve pressure on the spinal cord while
2
maintaining the stabilizing effects of the posterior elements of the vertebrae. Different
laminoplasty techniques have been described which vary by where the hinge and opening of the
lamina are developed. They can be broadly classified into two types namely open door (ODL)
and double door laminoplasty (DDL). Open door laminoplasty includes opening of the lamina
from either left or right side with the contralateral side acting as hinge. In double door
laminoplasty, the door is opened in the midline and the hinges are created bilaterally thereby
resulting in symmetric opening (Figure 3 and Figure 4).
Achieving and maintaining an increased sagittal diameter after cervical laminoplasty is
probably one of the most important criteria to facilitate neurological recovery. Laminar opening
can be maintained using sutures, autograft/allograft bone, hydroxyapatite (HA) spacers and
laminoplasty plates. The current study employs titanium plate and screw system for ODL and
HA spacers for DDL. The use of miniplates from cranial fixation systems for cervical
laminoplasties was first described by Frank et al. and later O’Brien et al. quantified the increased
spinal canal area using this technique [3, 4]. No intraoperative and postoperative complications
were observed [5]. A retrospective study performed on 104 patients with instrumented ODL
proved the technique to be a safe and reliable method for preventing progression of myelopathy
with multilevel involvement. There was no sign of fatigue failure or loosening of plates in this
large series of patients, thereby avoiding the revision for fixation failure which could have led to
the closure of lamina [6]. Shaffrey et al. performed a modified open door laminoplasty with
allograft bone and titanium miniplates and reported good short-term results in terms of
neurological outcome in younger patients with congenital spinal stenosis [7]. Hence, with the
increasing laminoplasty market, there is a need to evaluate the effect of these implants on the
biomechanical stability of the spine.
3
Compared to the open door laminoplasty, double-door laminoplasty has some theoretical
advantages like symmetrical expansion, and the potential for posterior fusion with a bone graft
bridge between the spinous process. It was reported that the group of patients with HA spacers
effectively maintained the range of motion compared to the group with autogenous iliac grafts
[8]. However, inadequate contact between the spacer and bone was one of the initial problems,
resulting in the instability of the spacer and restenosis of the spinal canal. During laminar
opening, the widened space was observed to be trapezoidal in both axial and frontal planes in a
finite element study conducted by Hirabayashi et al [9].The study also reported increased lateral
deviation of the spinous process with the hinge than without the hinge, and thereby suggesting
the importance of creation of hinge during laminar opening. Hence, most of the currently
available spacers are trapezoidal in shape to accommodate the contour of the spinous process.
Significance: The external and internal behavior of the spine after laminoplasty has been
of great interest to the clinicians and biomechanical researchers. Understanding the effect of this
surgical procedure on the biomechanics of the spine helps in better treatment and prevention of
spinal instability. To date, numerous clinical studies have been done to see the effect of multi-
level laminoplasty on the kinematics of spine. But, very few in vitro studies and to our
knowledge, this is the first computational study done to see the biomechanical response of the
spine to the surgical procedure. The current study emphasizes on simulating laminoplasty and
laminectomy in multi-level (C3-C6) finite element models. Moreover, the main aim of
laminoplasty is create a stable laminar arch to preserve the laminar opening. As hinge failure is a
commonly encountered problem during laminoplasty, it is necessary to understand the process of
bone remodeling post laminoplasty. Hence, the study also implements a computer simulation
method to predict bone remodeling in accordance with Wolff’s Law.
4
METHODS:
Specific Aim 1: Simulation of laminectomy and laminoplasty techniques for C3-C6 levels in
an intact C2-T1 finite element model.
In this study, a detailed validated subject-specific 3D finite element model of the cervical
spine (C2-T1) was used [10, 11]. The vertebral bodies were segmented from the CT images of
the cadaveric spine and were meshed with hexahedral elements using multi-block meshing
technique (IA-FeMesh). The vertebral body was divided into cortical and cancellous regions and
a Young’s modulus of 10GPa and 450MPa was assigned respectively. The model includes
clearly defined annulus and nucleus regions, five major spinal ligaments namely anterior
longitudinal ligament, posterior longitudinal ligament, ligamentum flavum, interspinous and
capsular ligaments. Additionally, the facet gap was modeled using the tabular pressure-
overclosure relationship in ABAQUS finite element software (Simulia, Providence, RI).
Validation of the model is essential but extremely difficult due to the varying material properties
along the length of the specimen. The current intact model was validated by comparing the finite
element predictions with subject-specific experimental data under similar loading and boundary
conditions. Subsequently, the intact C2-T1 finite element model was modified to simulate the
surgical procedure of laminectomy and laminoplasty at levels C3-C6; which are the most
commonly observed levels of stenosis. As an example, the procedure is explained clearly for C5
vertebrae through figures.
Simulation of Laminectomy: The lamina, spinous process and the associated ligaments
(interspinous ligaments, ligamentum flavum) were removed. Facet joints were kept intact
(Figure 2).
5
Simulation of ODL: A bicortical cut was simulated along the junction of the lamina and
the lateral mass of the intact vertebral mesh by completely removing a layer of elements. On the
contralateral side, a hinge of approximately 3-4mm was created along the junction of the lamina
and lateral mass by removing elements representing the unicortical layer (Figure 3(a)). The
spinous processes of the involved vertebrae (C3-C6) along with the interspinous ligaments were
resected. Additionally, the ligamentum flavum at the unaltered levels (C2-C3 and C6-C7) was
also cut partially on the open side of the lamina to allow laminar opening. Two screw holes in
the lateral mass and one hole in the lamina were created based on the desired plate position. Care
was taken to angle the screws away from the facet joints. The lamina of each vertebra, C3-C6,
was opened towards the hinge by applying a uniform load (Figure 3(b)). The laminar opening
space which is the distance between the lateral mass and the cut end of the lamina can vary from
8 to 12mm depending on the patient’s compressive pathology. We chose a laminar opening of
10mm for all the levels C3-C6. Our Previous studies on single level laminoplasty have shown an
increase of 56% in spinal canal area and 35% in sagittal canal diameter at C5 level [12]. The
stresses developed in the vertebra while opening the lamina tend to close it back and hence have
to be stabilized using plates and screws (Figure 3(c)). The titanium (Ti) plate and screws were
meshed with hexahedral elements using IA-FEMesh and were assigned an elastic modulus of
116GPa and Poisson’s ratio of 0.3. Small sliding contact was formulated at the interface between
the bone and the laminoplasty constructs. Additionally, the surfaces of the bone/screw and the
screw/plate were tied during the analysis.
Simulation of DDL: Two bilateral hinges at the junction of lateral mass and lamina were
created by removing the elements corresponding to the unicortical layer. Hinges should not be
too medial or too lateral as too medial may result in insufficient enlargement of the spinal canal,
6
and too lateral makes opening the split lamina difficult and may also cause facet fusion. The
spinous process is sagitally split until a laminar opening of 10mm is obtained. Laminar opening
is defined as the distance between the basal part of spinous process. The technique also involves
resection of interspinous ligaments at all the involved levels and partial removal of the
ligamentum flavum at the midline from C2 to C7 to allow for laminar opening. A 10mm
trapezoidal shaped hydroxyapatite (HA) spacer, meshed with hexahedral elements was used to
stabilize the lamina in the open position. The bony union observed in many clinical studies
between the HA spacers and spinous process was simulated by using the TIED command in
ABAQUS [13].
Figure 6 shows the simulated laminoplasty and laminectomy models.
Flexibility Test: The intact and all three surgically simulated models (namely ODL, DDL
and laminectomy) were tested using a pure moment of 2Nm under physiologic flexion/extension
(±MX), right/left lateral bending (±MZ), and right/left axial rotation (±MY) modes. The inferior
nodes of the T1 vertebra were fixed in all directions and a moment of 2Nm was applied on the
superior surface of C2. The analysis was performed using the finite element software ABAQUS
6.9. Ranges of motion, stresses in the annular regions of the intervertebral discs and cortical
regions of the vertebral bodies were analyzed and compared to that of the intact model.
Specific Aim 2: Application of adaptive bone remodeling theory to cervical laminoplasty
Literature presents several attempts made by various authors to quantify the bone
remodeling process. The theory of adaptive elasticity was developed to predict the behavior of
bone from one loading configuration to another by assuming a site-specific homeostatic
equilibrium strain state. Weinans et al. postulated the rate of change of apparent density of the
bone at a particular location in terms of mechanical stimulus like strain energy density (SED)
7
[14]. The difference between the actual SED (i.e. S), and the homeostatic SED (i.e.S0) is the
driving force for bone remodeling. As proposed by Carter et al., a lazy zone was introduced so
that a certain threshold level is exceeded before the bone reacts [15]. We chose to use a lazy zone
of 35% based on the earlier remodeling studies done on spine [16]. Equation1 describes the
nonlinear relationship between the mechanical stimulus and rate of change of density featuring
the lazy zone (1±s) that represents the minimum effective strain.
B(S-(1+s) S0) S> (1+s) S0
0 (1-s) S0 ≤ S ≤ (1+s) S0
B(S-(1-s) S0) S> (1+s) S0
B: Remodeling Rate s: Half width of lazy zone interval
S: Remodeling stimulating signal S0: Reference stimulating signal
Assuming that the behavior of remodeling is similar for all the altered levels, we chose to
implement remodeling at the C5 vertebra only. The remodeling code was written in FORTRAN
90. The integration between the remodeling code and finite element model is shown in Figure 5.
The objective function is defined as F/N; i 0,i
F S 1 s S │ │
N=Number of elements satisfying the objective criteria
Intact Model: Intact finite element model of C2-T1 was taken and a compressive load of
50N was applied on the top of the C2 surface. Subsequently, the bone remodeling algorithm was
applied to C5 vertebrae to determine optimal bone density distribution. The reference value (S0)
for each part was determined as the average SED value of all elements under compression.
Hence S0 is constant for all elements. If the objective criterion was satisfied, the material
properties were updated and the algorithm was repeated until convergence was met.
Laminoplasty Model: C2-T1 open door laminoplasty model was taken and bone
remodeling algorithm was implemented at C5 only. The site-specific SED values from the
SEDS
(1) d
dt
8
optimized intact model serve as the reference SED values for the laminoplasty model. Hence S0
is specific for each element. Hinge failure was simulated by removing a single row of elements
where maximum stress concentration occurred. The model was run under a 50N compressive
load and remodeling was performed until convergence was met.
RESULTS:
As mentioned earlier, the main aim of the laminoplasty technique is to hold the lamina in
the open position without failure to successfully decompress the spinal cord. The laminoplasty
constructs namely the screws and plates of ODL and the HA spacer of DDL were evaluated for
their stability. During all the six loading modes, the von Mises stresses in the plate and screw
were within the yield strength of the Ti implant i.e.917MPa.
Figure 7 compares the percent changes in the total range of motion (C2-T1) after
laminoplasty and laminectomy. The surgical procedures had significant effect in flexion while
minimal changes were observed in the other loading modes. ODL resulted in a 5.4% increase in
C2-T1 range of motion during flexion, while less than a 5% change was observed in the other
loading modes. DDL resulted in a significant 20% increase in the overall range of motion during
flexion with marginal changes in other loading modes. Laminectomy resulted in a substantial
57.5% increase in the total range of motion during flexion only with minimal changes in other
directions.
Considering the change in intersegmental rotations during flexion: After ODL, there was
a 39% and 20% increase in the motion at C2-C3 and C6-C7, while no substantial changes were
observed at the altered levels. The percent increase in the motion after DDL varied from 4.3% to
34.6%. Laminectomy at C3-C6 led to significant increase in motion across all the levels from
9
C2-C3 to C6-C7. These changes in the intersegmental rotations could be attributed to the
ligament resection at those levels.
A significant increase in the cortical stresses of the vertebral bodies was observed at
levels C3-C6 after the surgical procedures, with the posterior regions demonstrating higher
stresses than the anterior regions. DDL recorded higher stresses than ODL. This could be due to
the increased stresses developed in the vertebrae with opening both the lamina as opposed to
single lamina in ODL. Figure 8 shows the percent changes in the annular stresses of the
intervertebral discs after the simulated surgical procedures during flexion. Significant changes
were observed during flexion and correlated to the changes in the intersegmental rotation where
unaltered levels were affected after ODL and altered levels showed substantial increases after
DDL and laminectomy.
Figure 9(a) shows the density distribution for an optimal intact configuration under
compressive load. It resulted in the increase of the density values in the posterosuperior, mid-
inferior cortical and pedicle regions. These changes in the internal bone density can lead to
changes in the external morphology like thickening of the endplates in those regions. Figure 9(b)
shows the density distribution in a converged C5 vertebra after ODL. Since, the model was run
only under compression, most of the posterior elements were unloaded and as a result no major
differences in the density distribution were observed. The model shows a slight increase in the
density values at the hinge region suggesting bone formation.
DISCUSSION and CONCLUSION:
Numerous clinical studies have addressed the influence of multi-level laminoplasty on
the kinematics of spine. To our knowledge, currently there exists no other computational study
that compares the biomechanical effects of the two laminoplasty techniques. The results from the
10
current finite element predictions suggest the preservation of range of motion after open door
laminoplasty that corresponds well with the previous in vitro studies [17]. Kubo et al. showed an
inclination towards increased motion at all levels, which was also observed in the current study
after DDL, thereby explaining the role of lamina-ligamentum flavum complex in the stability of
spine [18]. As hypothesized, laminectomy resulted in a significant increase in motion during
flexion while no significant differences were observed in other loading modes as no facet injury
was simulated in the current study. The significant increase in the cortical stresses of the
vertebral body after both the laminoplasty and laminectomy procedures may be clinically
correlated to the degenerative process. To further augment/validate the finite element studies, we
are currently performing flexibility tests on cadaveric cervical specimens after laminoplasty and
laminectomy.
The greatest density changes in the optimized intact vertebrae occurred posteriorly and
mainly in the superior and inferior endplate regions. This is in accordance with the work done by
other authors, where the center of the endplate and the region between the endplates were less
dense compared to other regions of the vertebrae [16, 19]. It was observed that compression did
not have any major affect on the posterior elements that were significantly altered by
laminoplasty. Hence, we are currently extending bone remodeling to other loading conditions
like flexion and extension to get a better representation of density changes. It can be concluded
that adaptive bone remodeling theory can be applied to understand the process of bone
remodeling especially at the hinge region following surgical simulations like open door
laminoplasty.
11
Figures
Figure 1: Anatomy of Cervical spine Figure 2: Laminectomy on C5 vertebrae
(a) (b) (c)
OPEN DOOR LAMINOPLASTY: Figure 3 (a): C5 vertebrae showing hinge and open
side of the lamina (b): C5 vertebrae showing the Laminar opening (c) C5 vertebra stabilized with
plates and screws
DOUBLE DOOR LAMINOPLASTY Figure5: Bone Remodeling Algorithm
Figure 4: C5 vertebra stabilized with HA spacer
Lateral Mass
Hinge Bicortical cut
Hinge Hinge
Spinal Canal
12
(a) (b) (c)
Figure 6: C2-T1 Models (a) Laminectomy Model (b) Open Door Laminoplasty (c) Double Door
Laminoplasty
Figure 7: Percent changes in C2-T1 range of motion Figure 8: Percent changes in the
after ODL, DDL and laminectomy annular stresses during flexion
(a) (b)
Figure 9: (a) Density distribution in an optimized intact C5 vertebra (b) Density
distribution in a converged C5 vertebra after ODL
13
References
[1] Simeone, F. A., and Rothman, R., 1982, "Cervical Disc Disease," The Spine, pp. 440.
[2] Matsunaga, S., Sakou, T., and Nakanisi, K., 1999, Spinal Cord, 37(1) pp. 20-24.
[3] Frank, E., and Keenen, T. L., 1994, British Journal of Neurosurgery, 8(2) pp. 197-199.
[4] O'Brien, M. F., Peterson, D., Casey, A. T. H., 1996, Spine, 21(4) pp. 474.
[5] Deutsch, H., Mummaneni, P. V., Rodts, G. E., 2004, Journal of Spinal Disorders &
Techniques, 17(4) pp. 317.
[6] Dimar II, J. R., Bratcher, K. R., CCRP, D. C. B., 2009, American Journal of Orthopedics
(Belle Mead, NJ), 38(7) pp. E123.
[7] Shaffrey, C. I., Wiggins, G. C., Piccirilli, C. B., 1999, Journal of Neurosurgery: Pediatrics,
90(2).
[8] Ueyama, K., Harata, S., Okada, A., 2003, Cervical Laminoplasty, pp. 111.
[9] Hirabayashi, S., and Kumano, K., 2000, Journal of Musculoskeletal Research, 4(1) pp. 47-54.
[10] Kallemeyn, N., Gandhi, A., Kode, S., 2010, Medical Engineering & Physics, 32(5) pp. 482-
489.
[11] Wheeldon, J. A., Pintar, F. A., Knowles, S., 2006, Journal of Biomechanics, 39(2) pp. 375-
380.
[12] Srinivas, T., C., Nicole, K.,A., Kiran, S.,H., 2008, Proceedings of the ASME 2008 Summer
Bioengineering Conference.
[13] Iguchi, T., Kanemura, A., Kurihara, A., 2003, Journal of Neurosurgery: Pediatrics, 98(2) .
[14] Weinans, H., Huiskes, R., and Grootenboer, H., 1992, Journal of Biomechanics, 25(12) pp.
1425-1441.
[15] Carter, D. R., 1984, Calcified Tissue International, 36pp. 19-24.
[16] Grosland, N. M., and Goel, V. K., 2007, Spine, 32(23) pp. E667.
[17] Subramaniam, V., Chamberlain, R. H., Theodore, N., 2009, Spine, 34(16) pp. E573.
[18] Kubo, S., Goel, V. K., Yang, S. J., 2003, Spine, 28(3) pp. 227.
[19] Grant, J. P., Oxland, T. R., and Dvorak, M. F., 2001, Spine, 26(8) pp. 889.